• Vimal S Gunasekaran, Dustin Hejdak, Benjamin Meyer, Andrew Klein, and Kevin Koch.
• Department of Radiology, Medical College of Wisconsin, 8701 W Watertown Plank Rd, Wauwatosa, WI, 53226, USA. vgunasekaran@mcw.edu.
• Eur Spine J. 2021 Nov 1; 30 (11): 3319-3323.

PurposeClinical evaluation of lumbar foraminal stenosis typically includes qualitative assessments of perineural epidural fat content around the spinal nerve root and evaluation of nerve root impingement. The present study investigates the use of several morphological MRI-derived metrics as quantitative predictors of foraminal stenosis grade.Methods62 adult patients that underwent lumbar spine MRI evaluation over a 1-month duration in 2018 were included in the analysis. Radiological gradings of stenosis were captured from the existing clinical electronic medical record. Clinical gradings were recorded using a 0-5 scale: 0 = no stenosis, 1 = mild stenosis, 2 = mild-moderate stenosis, 3 = moderate stenosis, 4 = moderate-severe stenosis, 5 = severe stenosis. Quantitative measures of perineural epidural fat volume, nerve root cross-sectional area, and lumbar pedicle length were derived from T1 weighted sagittal spine MRI on each side of all lumbar levels. Spearman correlations of each measured metric at each level were then computed against the stenosis gradings.ResultsA total of 347 volumetric segmentation and radiological foraminal stenosis grade sets were derived from the 62-subject study cohort. Statistical analysis revealed significant correlations (p < 0.001) between the volume of perineural fat and stenosis grades for all lumbar vertebral levels.ConclusionThe results of the study have demonstrated that segmented volumes of perineural fat predict the severity of clinically scored foraminal stenosis. This finding motivates further development of automated perineural fat segmentation methods, which could offer a quantitative imaging biometric that yields more reproducible diagnosis, assessment, and tracking of foraminal stenosis.© 2021. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

15 keywords...

hide…